llama : use n_embd_head_v instead of n_embd_head_k when reshaping kqv #7327
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
mofosyne
added
bugfix
…d n_embd_head_k when making a view of cached value vectors.
Collaborator
Author
|
I found another place when variables for key vectors were used for processing value vectors, so I added another commit to this PR. |
ggerganov
approved these changes
May 17, 2024
Member
ggerganov
left a comment
ggerganov
left a comment
There was a problem hiding this comment.
Which models are affected by this?
Collaborator
Author
|
DeepSeek-V2 needs this since it has n_embd_head_k != n_embd_head_v, I'm not sure about other models: |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
When reshaping kqv at the end of llm_build_kqv() n_embd_head_k is incorrectly used instead of n_embd_head_v to calculate kqv dimensions.